1: \documentclass[12pt]{article} \pagestyle{plain}
2: \setlength{\textheight}{23cm} \setlength{\textwidth}{16.7cm}
3: \setlength{\columnsep}{2.0pc} \setlength{\topmargin}{-0.2in}
4: \setlength{\headheight}{0.0in} \setlength{\headsep}{0.0in}
5: \setlength{\oddsidemargin}{0.1in} \setlength{\parindent}{1pc}
6: \usepackage{graphicx}
7: \usepackage{epsfig}
8: \usepackage{amsbsy}
9:
10: \begin{document}
11: \baselineskip 10mm
12:
13: \centerline{\large \bf Decay and fusion as two different
14: mechanisms} \centerline{\large \bf of stability loss for the
15: (C$_{20}$)$_2$ cluster dimer}
16:
17: \vskip 6mm
18:
19: \centerline{L. A. Openov$^{*}$ and A. I. Podlivaev}
20:
21: \vskip 4mm
22:
23: \centerline{\it Moscow Engineering Physics Institute (State
24: University), 115409 Moscow, Russia}
25:
26: \vskip 2mm
27:
28: $^{*}$ E-mail: LAOpenov@mephi.ru
29:
30: \vskip 8mm
31:
32: \centerline{\bf ABSTRACT}
33:
34: The thermal stability of the (C$_{20}$)$_2$ cluster dimer
35: consisting of two C$_{20}$ fullerenes is examined using a
36: tight-binding approach. Molecular dynamics simulations of the
37: (C$_{20}$)$_2$ dimer at temperatures $T=$ 2000 - 3500 K show that
38: the finite lifetime $\tau$ of this metastable system is determined
39: by two fundamentally different processes, the decay of one of the
40: C$_{20}$ fullerenes and the fusion of two C$_{20}$ fullerenes into
41: the C$_{40}$ cluster. The activation energies for these processes
42: $E_a\approx$ 3.4 and 2.7 eV, respectively, as well as their
43: frequency factors, have been determined by analyzing the
44: dependence of $\tau$ on $T$.
45:
46: \newpage
47:
48: In 2000, after the experimental detection of the smallest
49: fullerene C$_{20}$ [1], Fig. 1a, on the surface of which C-C bonds
50: form only pentagons, whereas hexagons are absent (in contrast to
51: C$_N$ fullerenes with $N > 20$), a problem of synthesizing the
52: condensed phase (fullerite) based on this fullerene by analogue
53: with the known fullerite based on C$_{60}$ fullerenes [2, 3]
54: arised. Interest in this problem is stimulated, because the
55: C$_{20}$ fullerite may be a high-temperature superconductor [4-6].
56: Indeed, preliminary calculations point to a possibility of
57: formation of three-dimensional crystals consisting of C$_{20}$
58: fullerenes [4-6]. There is yet disagreement on the crystal
59: structure of such a hypothetical cluster matter [5,6], but all
60: authors agree that C$_{20}$ clusters in it should be connected
61: with each other by strong covalent bonds. This is a fundamental
62: difference of the C$_{20}$ fullerite from the C$_{60}$ fullerite
63: in which clusters are bonded with each other due to a weak van der
64: Waals attraction.
65:
66: The strong intercluster interaction can significantly distort the shape of the C$_{20}$ clusters in the fullerite with
67: respect to the spherical shape of the C$_{20}$ fullerene and, as a consequence, result in the transition to another
68: crystalline (or even amorphous) modification. It seems to be reasonable to begin the analysis of the stability of the
69: C$_{20}$ fullerite with the analysis of the stability of the simplest system consisting of two interacting C$_{20}$
70: fullerenes, the (C$_{20}$)$_2$ cluster dimer. Experimental indications of the existence of charged (C$_{20}$)$_2^+$
71: dimers [as well as complexes (C$_{20}$)$_k^+$with $k=$ 3 - 13] were obtained in [7]. The dimerization of neutral
72: fullerenes C$_{20}$ was theoretically studied in [4] and, most comprehensively, in [8], where eight different
73: (C$_{20}$)$_2$ isomers were found, their structural and energy characteristics were calculated, and it was shown that
74: the so-called {\it open}-[2+2] isomer has the lowest energy, see Fig. 1b (numbers in the square brackets are the numbers
75: of the atoms of each fullerene that are involved in the formation of intercluster bonds).
76:
77: The aim of this work is to theoretically analyze the stability of
78: the (C$_{20}$)$_2$ dimer with respect to its transition to another
79: atomic configuration. It is known (see, e.g., [9]) that the
80: specific (per atom) binding energy of carbon clusters, which is
81: defined as the difference between the energy of an isolated carbon
82: atom and the specific potential energy of the
83: $N$-atomic cluster, $E_b(N)/N=E(1)-E_{pot}(N)/N$, increases with
84: $N$ (this is the property of the overwhelming majority of the
85: elements of the Periodic Table, except for metastable clusters of
86: nitrogen, helium, and some other elements [10, 11]). For this
87: reason, it is energetically favorable for two (or several) carbon
88: clusters to merge into one larger cluster, thereby increasing the
89: binding energy [10] (i.e., reducing the potential energy), by
90: analogy with fusion of light atomic nuclei. At the same time, our
91: previous analysis [12] of the thermal stability of the isolated
92: C$_{20}$ fullerene showed that, when heated, it passes to various
93: energetically unfavorable configurations with lower $E_b$ values
94: rather than to a more stable isomer with the maximum binding
95: energy ("bowl"). This is due to the ability of carbon structures
96: to form numerous “intermediate” metastable states separated by
97: high energy barriers from configurations with a lower potential
98: energy.
99:
100: Since the C$_{20}$ fullerenes in the (C$_{20}$)$_2$ dimer are, on
101: the one hand, at small distances from each other and hence strongly
102: interact with each other and, on the other hand, both preserve
103: the overall spherical shape (see Fig. 1b), the (C$_{20}$)$_2$ dimer
104: can loose its “molecular” structure both through the fusion of two
105: C$_{20}$ fullerenes to the C$_{40}$ cluster and through the decay
106: of one of them [12]. A question arises of which of two factors,
107: thermodynamic (the minimum of the energy of the final product
108: after fusion) or kinetic (the thermally activated transition to
109: the metastable configuration with a high energy upon decay),
110: finally determines the stability of the (C$_{20}$)$_2$ dimer? To
111: answer this question, we study the dynamics of the (C$_{20}$)$_2$
112: dimer numerically at high (sufficient to overcome the
113: corresponding potential barriers $U\sim 5$ eV [12]) temperatures.
114:
115: To calculate the energies of arbitrary atomic configurations, we
116: used the tight binding model [13], which is a reasonable
117: compromise between {\it ab initio} calculations and too simplified
118: empirical approaches with classical inter-atomic potentials. This
119: model well describes both small clusters and macroscopic carbon
120: structures. Previously, we used it to simulate the fullerenes
121: C$_{20}$, C$_{60}$, and other carbon systems [10, 12, 14-18]. For
122: the binding energy of an isolated C$_{20}$ fullerene, this method
123: gives $E_b=6.08$ eV/atom [12] (in agreement with a value of
124: $E_b=6.36$ eV/atom obtained in {\it ab initio} calculations [19).
125: For the binding energy of the {\it open}-[2+2] isomer of the
126: (C$_{20}$)$_2$ dimer, see Fig. 1b, this method gives a value of
127: $E_b=6.20$ eV/atom (which corresponds to the coagulation energy of
128: two fullerenes $\Delta
129: E=2E_{pot}$[C$_{20}$]-$E_{pot}$[(C$_{20}$)$_2$] = 4.9 eV, to be
130: compared with the value $\Delta E=6.3$ eV obtained by the density
131: functional method [8]).
132:
133: To analyze the thermal (it can also be called kinetic) stability of the (C$_{20}$)$_2$ dimer, we study its dynamics
134: numerically for the initial temperature $T_{ini}=$ 2000 - 3500 K and fixed total energy $E_{pot}+E_{kin}=$ const, where
135: $E_{kin}$ is the kinetic energy in the center-of-mass syatem. In this formulation of the problem, the system temperature
136: $T$ is a measure of the energy of the relative motion of atoms (for more details, see [12, 17, 18]). The values of $T$
137: were calculated by the formula $\langle E_{kin}\rangle = \frac{1}{2}k_BT(3N-6)$, where $k_B$ is the Boltzmann
138: constant, $N=40$ is the number of atoms in the (C$_{20}$)$_2$ dimer, and the angular brackets stand for averaging
139: over $10^3 - 10^4$ steps of the molecular dynamics simulation (the time of one step is $t_0=2.72\cdot10^{–16}$ s).
140:
141: We observed two fundamentally different types of the $T(t)$
142: dependence. They are shown in Fig. 2. In the first case (Fig. 2a),
143: the dimer temperature at $t\approx 4.4$ ns drops stepwise by
144: $\Delta T\approx$ 200 K. The decrease in temperature points to a
145: decrease in the kinetic energy and, hence, an increase
146: in the potential energy (because $E_{pot}+E_{kin}=$ const).
147: Therefore, the (C$_{20}$)$_2$ dimer passes to a less favorable
148: configuration. Our analysis showed that this transition is caused
149: by the decay of one of the C$_{20}$ fullerenes, whereas the other
150: fullerene preserved the cage shape, see Fig. 3a. The binding
151: energy of the metastable state formed upon the decay is equal to
152: (after the relaxation to the local minimum of the potential
153: energy) $E_b=6.14$ eV/atom, which is lower than the value of $E_b$
154: for the initial (C$_{20}$)$_2$ dimer by $\Delta E_b=0.06$ eV/atom.
155: This $\Delta E_b$ value is consistent with the value $\Delta
156: T\approx$ 200 K, taking into account that, in thermal oscillations
157: near the local minimum of the potential energy $E_{pot}^{min}$ ,
158: the average kinetic energy of the system is approximately (up to
159: anharmonic corrections) equal to the average increase in the
160: potential energy $\Delta E_{pot}=E_{pot}-E_{pot}^{min}$, so that,
161: since the total energy is constant, i.e.,
162: $E_{pot}+E_{kin}=E_{pot}^{min}+\Delta E_{pot}+E_{kin}\approx
163: E_{pot}^{min}+2E_{kin}\approx E_{pot}^{min}+3Nk_BT$, in the
164: $1\rightarrow 2$ transition from the vicinity of one local minimum
165: to the vicinity of the other local minimum, the relation
166: $E_{pot}^{min,1}+3Nk_BT_1\approx E_{pot}^{min,2}+3Nk_BT_2$ is
167: valid, from which it follows that $\Delta T=T_1-T_2\approx \Delta
168: E_b/3k_B$ for $E_{pot}^{min,2}>E_{pot}^{min,1}$ taking into
169: account the definition of $E_b$.
170:
171: On contrary, for the second case (Fig. 2b), the temperature at
172: $t\approx 4.4$ ns increases stepwise by $\Delta T\approx 200$ K.
173: This indicates an increase in the kinetic energy, i.e., a decrease
174: in the potential energy. Therefore, the system passes to the
175: energetically favorable configuration. However, since the initial
176: {\it open}-[2+2] isomer has the minimum energy among all
177: (C$_{20}$)$_2$ dimers, this transition indicates that the C$_{20}$
178: fullerenes loss their individuality and merge into the C$_{40}$
179: cluster. We analized the atomic structure after the temperature
180: jump and showed that this is indeed the case. The C$_{40}$ cluster
181: formed upon fusion is shown in Fig. 3b. It can be considered as
182: one of nonclassical, strongly imperfect isomers of the C$_{40}$
183: fullerene: the C-C bonds in it form not only pentagons and
184: hexagons, but also two 9-member rings. The binding energy of this
185: cluster after relaxation, $E_b=6.25$ eV/atom, is higher than the
186: value for the (C$_{20}$)$_2$ dimer by $\Delta E_b=0.05$ eV/atom,
187: which is consistent with a value of $\Delta T\approx 200$ K upon
188: fusion.
189:
190: We also observed the fusion of the C$_{20}$ fullerenes into
191: isomers with higher binding energies. Fig. 4 shows the $T(t)$
192: dependence for the case, where an imperfect isomer of the C$_{40}$
193: fullerene with one heptagon and $E_b=6.49$ eV/atom is formed upon
194: fusion at $t\approx 0.34$ ns. At $t\approx 0.46$ ns, the
195: annealing of the defect takes place and the C$_{40}$ fullerene
196: with $E_b=6.53$ eV/atom is formed, on the surface of which the C-C
197: bonds form 12 pentagons and 10 hexagons, see Fig. 5.
198: In this case, the temperature increases
199: sequentially by $\Delta T\approx 1100$ and 150 K, which is
200: consistent with changes $\Delta E_b$, see above. Transitions of
201: the (C$_{20}$)$_2$ dimer into other isomers of the C$_{40}$
202: fullerene were also observed. The general statistics is as follows
203: ($N_k$ is the number of $k$-gons with $k > 6$ in the C$_{40}$
204: isomer). Among 57 sets of the initial displacements and velocities
205: of atoms that correspond to different initial
206: temperatures $T_{ini}$, the fusion into the C$_{40}$ cluster with
207: $N_9=2$ occurs most frequently (16 times), see Fig. 3b. In
208: addition, the clusters with $N_{10}=2$ and $N_8=1$ (six times);
209: $N_9=N_8=N_7=1$ (once); $N_8=2$ (twice); $N_8=N_7=1$ (once);
210: $N_7=1$ (once); and $N_7=1$ (once) are formed. In other 29 cases,
211: the decay of one of the C$_{20}$ fullerenes takes place (Fig. 3a). The
212: further evolution of the system after it looses its molecular
213: structure is as follows. The decay of one of the fullerenes in the
214: (C$_{20}$)$_2$ dimer is followed by the decay of the second
215: fullerene, while fusion of two fullerenes into the C$_{40}$
216: cluster is followed by the decay of this cluster, so that the
217: final product is always a quasi-two-dimensional or
218: quasi-one-dimensional configuration as in the decay of the
219: isolated C$_{20}$ fullerene [12]. We never observed the return of
220: the system to the initial {\it open}-[2+2] configuration. The
221: fission of the (C$_{20}$)$_2$ dimer into two isolated C$_{20}$
222: fullerenes has not been observed as well.
223:
224: There is correlation between the value of $T_{ini}$ and the
225: mechanism of the loss of the stability of the (C$_{20}$)$_2$
226: dimer. The loss of the molecular structure of the dimer occurs
227: primarily due to fusion of two C$_{20}$ fullerenes into the
228: C$_{40}$ cluster at $T_{ini} < 2500$ K, due to the decay of one of
229: the fullerenes at $T_{ini} > 3000$ K, and due to both those
230: processes at 2500 K $< T_{ini} < 3000$ K, see Fig.6. To determine
231: the activation energies $E_a$ for the decay and fusion, we
232: analyzed the temperature dependence of the lifetime $\tau$ of the
233: (C$_{20}$)$_2$ dimer, using the Arrhenius formula with the
234: finite-heat-bath correction [18, 20, 21]
235: \begin{equation}
236: \tau^{-1}(T)=A\cdot\mathrm{exp}\left[-\frac{E_{a}}{k_{B}T\left(1-E_{a}/2CT\right)}\right].
237: \end{equation}
238: Here, $A$ is the frequency factor, $C$ is the microcanonical
239: specific heat, which we set to $C=(3N-6)k_B$, where $N=40$ (as
240: will be shown below, $E_a/2CT\approx 0.05 << 1$ for $T\approx
241: 3000$ K). Figure 6 shows the calculated logarithm of $\tau$ as a
242: function of ($T_{ini}^{*}$)$^{-1}$, where
243: $T_{ini}^{*}=T_{ini}(1-E_a/2CT_{ini})$, see Eq. (1). It is seen
244: that this dependence in a quite wide range $T_{ini} = 2000 - 3000$
245: K is given, in the first approximation, by the straight line for
246: both the decay and fusion processes; i.e., $A(T)\approx$ const
247: (since the spread of the points is sufficiently large,
248: particularly for fusion processes, it is impossible to reliably
249: determine the temperature dependence of $A$). Using the slopes of
250: these straight lines and the points of their intersection with the
251: ordinate axis, we determined the activation energies $E_a=3.4\pm
252: 0.2$ eV and $E_a=2.7\pm 0.3$ eV, as well as the frequency factors
253: $A=(1.8\pm 0.3)\cdot 10^{17}$ s$^{-1}$ and $A=(5.6\pm 1.1)\cdot
254: 10^{15}$ s$^{-1}$ for the decay and coalescence processes,
255: respectively. [Since $E_a$ enters both into the nominator of the
256: exponential in Eq. (1) and into the “renormalized” initial
257: temperature, it is determined self-consistently by successive
258: iterations]. The values found for $E_a$ are one fourth to one
259: third of the fragmentation activation energy of the C$_{60}$
260: fullerene (frequency factors are smaller by two to four orders of
261: magnitude) [18].
262:
263: In conclusion, we note that the value $E_a\approx 3.4$ eV for the decay of one C$_{20}$ fullerene in the (C$_{20}$)$_2$
264: dimer is approximately one half of the value $E_a\approx 7$ eV for the decay of the isolated C$_{20}$ fullerene [12].
265: This fact, along with the appearance of the second mechanism of the loss of the molecular structure (fusion of two C$_{20}$
266: fullerenes into the C$_{40}$ cluster) with a lower value of $E_a\approx 2.7$ eV, indicates that stability decreases
267: when the (C$_{20}$)$_2$ dimer is formed from two C$_{20}$ cages.
268: In three-dimensional cluster structures based on C$_{20}$ fullerenes, both the stabilization of the metastable state due to an
269: increase in the number of nearest neighbors of each C$_{20}$ fullerene and the appearance of new channels of
270: stability loss can be expected. We are going
271: to study this problem in the next work.
272:
273: \newpage
274:
275: \begin{thebibliography}{10}
276:
277: \bibitem{1}H.Prinzbach, A.Weller, P.Landenberger, F.Wahl, J.Worth, L.T.Scott,
278: M.Gelmont, D.Olevano, B. von Issendorff, Nature $\mathbf{407}$, 60
279: (2000).
280: \bibitem{2}H.W.Kroto, J.R.Heath, S.C.O'Brien, R.F.Curl, R.E.Smalley, Nature
281: $\mathbf{318}$, 162 (1985).
282: \bibitem{3}A.V.Eletskiy, B.M.Smirnov, Usp. Fiz. Nauk {\bf 165}, 977 (1995) [Phys. Usp. {\bf 38}, 935 (1995)].
283: \bibitem{4}Y.Miyamoto, M.Saito, Phys. Rev. B $\mathbf{63}$, 161401 (2001).
284: \bibitem{5}S.Okada, Y.Miyamoto, M.Saito, Phys. Rev. B $\mathbf{64}$, 245405
285: (2001).
286: \bibitem{6}I.Spagnolatti, M.Bernasconi, G.Benedek, Europhys. Lett. $\mathbf{59}$,
287: 572 (2002).
288: \bibitem{7}R.Ehlich, P.Landenberger, H.Prinzbach, J. Chem. Phys. $\mathbf{115}$,
289: 5830 (2001).
290: \bibitem{8}C.H.Choi, H.-I.Lee, Chem. Phys. Lett. $\mathbf{359}$, 446 (2002).
291: \bibitem{9}D. Tom{\fontencoding{T1}\selectfont\symbol{225}}nek, M.A. Schluter,
292: Phys. Rev. Lett. $\mathbf{67}$, 2331 (1991).
293: \bibitem{10}V.F.Elesin, N.N.Degtyarenko, L.A.Openov, Inzh. Fiz. {\bf 3}, 2 (2002), in Russian
294: \bibitem{11} V.F.Elesin, N.N.Degtyarenko, N.V.Matveev, A.I.Podlivaev, L.A.Openov,
295: Zh. Eksp. Teor. Fiz. {\bf 128}, 54 (2005) [JETP {\bf 101}, 44 (2005)].
296: \bibitem{12} I.V.Davydov, A.I.Podlivaev, L.A.Openov, Fiz. Tverd. Tela (St. Petersburg) {\bf 47}, 751 (2005)
297: [Phys. Solid State {\bf 47}, 778 (2005)]; cond-mat/0503500.
298: \bibitem{13}C.H. Xu, C.Z. Wang, C.T. Chan, K.M. Ho, J. Phys.: Condens. Matter
299: $\mathbf{4}$, 6047 (1992).
300: \bibitem{14}L.A.Openov, V.F.Elesin, Pis’ma Zh. Eksp. Teor. Fiz. {\bf 68}, 695 (1998) [JETP Lett. {\bf 68}, 726 (1998)];
301: physics/9811023.
302: \bibitem{15}V.F.Elesin, A.I.Podlivaev, L.A.Openov, Phys. Low-Dim. Struct. $\mathbf{11/12}$,
303: 91 (2000); physics/0104058.
304: \bibitem{16}N.N.Degtyarenko, V.F.Elesin, N.E.L’vov, L.A.Openov, A.I.Podlivaev, Fiz. Tverd. Tela (St. Petersburg)
305: {\bf 45}, 954 (2003) [Phys. Solid State {\bf 45}, 1002 (2003)].
306: \bibitem{17}A.I.Podlivaev, L.A.Openov, Pis’ma Zh. Eksp. Teor. Fiz. {\bf 81}, 656 (2005)
307: [JETP Lett. {\bf 81}, 533 (2005)]; cond-mat/0506571.
308: \bibitem{18}L.A.Openov, A.I.Podlivaev, Pis’ma Zh. Eksp. Teor. Fiz. {\bf 84}, 73 (2006)
309: [JETP Lett. {\bf 84}, 68 (2006)]; cond-mat/0609267.
310: \bibitem{19}R.O.Jones, G.Seifert, Phys. Rev. Lett. $\mathbf{79}$, 443 (1997).
311: \bibitem{20}C.E.Klots, Z. Phys. D $\mathbf{20}$, 105 (1991).
312: \bibitem{21}J.V.Andersen, E.Bonderup, K.Hansen, J. Chem. Phys. $\mathbf{114}$,
313: 6518 (2001).
314:
315:
316: \end{thebibliography}
317:
318: %\newpage
319:
320: %\centerline{Figure captions}
321:
322: \newpage
323:
324: \includegraphics[width=\hsize,height=19cm]{Fig1.eps}
325:
326: \vskip 5mm
327:
328: Fig. 1. (a) C$_{20}$ fullerene with the binding energy $E_b=6.08$ eV/atom and (b) (C$_{20}$)$_2$
329: cluster dimer with the binding energy $E_b=6.20$ eV/atom.
330:
331: \newpage
332:
333: \includegraphics[width=10cm,height=17cm]{Fig2.eps}
334:
335: \vskip 5mm
336:
337: Fig. 2. Time dependences of the temperature $T$ of the (C$_{20}$)$_2$ cluster dimer obtained from the molecular
338: dynamics simulation by averaging the kinetic energy over each $10^4$ steps. The total energy of the system is constant.
339: The time of one step is equal to $t_0=2.72\cdot10^{-16}$ s. The initial temperature is $T_{ini}=2415\pm 10$ K (a) and
340: $2285\pm 10$ K (b).
341:
342: \newpage
343:
344: \includegraphics[width=13cm,height=18cm]{Fig3.eps}
345:
346: \vskip 5mm
347:
348: Fig. 3. (a) Atomic configuration after the decay of one of the C$_{20}$ fullerenes in the (C$_{20}$)$_2$ dimer, see Fig. 2a.
349: The binding energy after relaxation is $E_b=6.14$ eV/atom. (b) The C$_{40}$ cluster formed upon fusion of two C$_{20}$
350: fullerenes in the (C$_{20}$)$_2$ dimer, see Fig. 2b. The binding energy after relaxation is $E_b=6.25$ eV/atom.
351:
352: \newpage
353:
354: \includegraphics[width=15cm,height=14cm]{Fig4.eps}
355:
356: \vskip 5mm
357:
358: Fig. 4. Same as in Fig. 2, for $T_{ini}=2820\pm 10$ K.
359:
360: \newpage
361:
362: \includegraphics[width=13cm,height=18cm]{Fig5.eps}
363:
364: \vskip 5mm
365:
366: Fig. 5. (a) Defect isomer of the C$_{40}$ fullerene that is formed
367: upon fusion of two C$_{20}$ fullerenes in the (C$_{20}$)$_2$
368: dimer, see Fig. 4. The binding energy after relaxation is
369: $E_b=6.49$ eV/atom. (b) The C$_{40}$ fullerene formed after
370: annealing of the imperfect isomer. The binding energy after
371: relaxation is $E_b=6.53$ eV/atom. The distant atoms are not shown
372: for clarity.
373:
374: \newpage
375:
376: \includegraphics[width=16cm,height=15cm]{Fig6.eps}
377:
378: \vskip 5mm
379:
380: Fig. 6. Logarithm of the lifetime $\tau$ (in seconds) of the (C$_{20}$)$_2$ cluster dimer for the processes of the decay
381: of one of C$_{20}$ fullerenes (circles) or fusion of two C$_{20}$ fullerenes into the C$_{40}$ cluster (squares) versus the
382: inverse temperature (Kelvins), taking the finite-heat-bath correcrtion into account, see text. The solid and dashed lines
383: are the linear approximations for decay and fusion, respectively.
384:
385: \end{document}
386: